CAD file summary information automatic extraction method and device, medium and product

Abstract

The invention provides a CAD file summary information automatic extraction method and device, a medium and a product, and relates to the technical field of computer picture word process.The method comprises the steps that a target CAD file is analyzed through a CAD file analysis development tool, and first word information and geometric attribute information corresponding to the first word information are obtained; retrieving the first character information according to the retrieval keyword to obtain second character information; according to the corresponding geometric attribute information in the second text information, carrying out positioning adjustment display through a CAD file analysis development tool to obtain a picture file; calling a first large language model to perform character recognition and arrangement processing on the picture to form a text list; and calling the second large language model to obtain summary information of the target CAD file according to a preset prompt template and the text list. According to the scheme, when a large batch of CAD files are processed, the processing efficiency is remarkably improved.

Description

Technical Field

[0001] The present application relates to the technical field of computer image and text processing, and in particular to a method, device, medium and product for automatically extracting summary information from CAD files. Background Art

[0002] With the rapid development of the construction industry, CAD software has become a vital tool for architectural design. Standardized management of design files, particularly those in DWG format, is crucial in building project management. These files contain extensive design details and technical information, requiring effective archiving and management to facilitate project traceability and design reuse.

[0003] In practice, project managers primarily use two methods to process DWG files. The first is a purely manual method, where a technician directly opens the DWG file, manually reviews the drawing content, including layout, annotations, and other information, manually writes a summary, and enters it into the project management system.

[0004] As the scale of construction projects continues to expand, a single project often contains hundreds or even thousands of DWG files. Manually processing these files would take up a significant amount of human resources. Furthermore, since the manual review and editing process is susceptible to factors such as fatigue and work efficiency, it is easy to miss important information or have inaccurate summary descriptions, affecting the quality and efficiency of project management. Furthermore, existing AI-assisted methods have low accuracy when recognizing text on drawings, requiring a significant amount of manpower for review and correction, resulting in low processing efficiency. Summary of the Invention

[0005] In order to solve the above technical problems, the present application provides a method, device, medium and product for automatically extracting CAD file summary information, which are used to solve the technical problem of low efficiency in CAD file summary extraction.

[0006] In a first aspect, the present application provides a method for automatically extracting summary information from a CAD file, comprising: The target CAD file is parsed using a CAD file parsing development tool to obtain first text information in the target CAD file, the first text information including a first field set and text geometric attribute information corresponding to each field in the first field set, the text geometric attribute information including text coordinates, text size, and text rotation angle; the first text information is searched based on a current search keyword to obtain second text information, the second text information including a second field set and the text geometric attribute information corresponding to each field in the second field set, wherein the second field set includes some fields in the first field set; based on the text geometric attribute information corresponding to each field in the second field set, the CAD file parsing development tool positions and adjusts each field in the second field set for display, to obtain an image file, wherein the image file includes an image corresponding to each field in the second field set; the first language model is invoked to perform text recognition on each image in the image file, and the recognition results are arranged to form a text list; the second language model is invoked based on a preset prompt template and the text list to obtain summary information of the target CAD file, wherein the preset prompt template is used to instruct to extract information associated with the search keyword from the text list and output it in a predetermined format.

[0007] Through the above technical solution, a CAD file parsing development tool is used to accurately locate and extract attributes from text information, obtain an actual image of the text in the CAD file, then process the actual image through the first language model, and generate summary information based on preset prompt words through the second language model. This achieves accurate recognition and semantic understanding of the text information in the CAD file, effectively solving the technical problems of direct extraction of CAD file text information in the existing technology, such as fragmentation, inaccurate extraction results, and low information extraction efficiency. It also achieves automated, intelligent, and standardized extraction of CAD file summary information. This solution ensures the accuracy of text information extraction through precise positioning of text geometric attribute information, ensures the integrity of the extracted content through step-by-step image recognition and semantic understanding, and ensures the uniformity of the output format through standardized guidance of preset prompt templates. Ultimately, it achieves the technical effects of improving the efficiency of CAD file information extraction, reducing manual processing costs, and enhancing the usability of extraction results.

[0008] Optionally, the step of searching the first text information according to the current search keyword to obtain the second text information specifically includes: The historical search keywords are analyzed through the third language model to adjust the current search keywords; the first text information is searched according to the adjusted search keywords to obtain the second text information matching the adjusted search keywords.

[0009] The above technical solution, which utilizes a historical search keyword analysis mechanism and dynamic keyword adjustment strategy based on the third language model, intelligently optimizes and supplements user-entered search keywords, effectively resolving technical issues in existing technologies such as single search keywords, low matching accuracy, and inaccurate search results. This solution significantly improves search efficiency and precisely matches search results. By analyzing historical search data and intelligently adjusting search keywords to expand the effective search range, the solution improves search accuracy through optimized keywords, ultimately achieving the technical effects of improving search efficiency, enhancing the relevance of search results, and enhancing the user search experience.

[0010] Optionally, the step of analyzing historical search keywords using the third language model and adjusting the current search keywords specifically includes: The third language model is called to perform cluster analysis on the historical search keywords to obtain multiple keyword categories, each of which includes a set of keywords with similar semantics; the current search keyword is matched with the multiple keyword categories for similarity to determine the target keyword category; based on the target keyword category and combined with the industry terminology knowledge base, the search keyword input by the user is supplemented and optimized to generate the adjusted search keyword.

[0011] By adopting the above technical solution, a keyword classification system based on semantic clustering and a supplementary optimization mechanism for the industry terminology knowledge base, the intelligent expansion and specialized adjustment of search keywords are achieved, effectively solving technical problems in existing technologies such as insufficient semantic understanding of keywords, incomplete coverage of professional terms, and a single search dimension, thereby achieving multi-dimensional optimization and professional improvement of search keywords. This solution establishes a keyword association network through semantic clustering, ensures the accuracy of optimization direction through similarity matching, and achieves precise supplementation of professional fields through the industry terminology knowledge base, ultimately achieving the technical effects of improving search coverage, enhancing search professionalism, and optimizing search results.

[0012] Optionally, the step of positioning, adjusting, and displaying each field in the second field set using the CAD file parsing and development tool according to the text geometric attribute information corresponding to each field in the second field set to obtain an image file specifically includes: The target field is any field in the second field set. For the target field, a corresponding target image is obtained in the following manner, and the image file includes the target image: using the CAD file parsing development tool, according to the text coordinates corresponding to the target field, the text area corresponding to the target field is located; according to the mapping relationship between the text size corresponding to the target field in the text area and the screen pixels, the screen display size of the target field is calculated; based on the screen display size and the text rotation angle corresponding to the target field, the display boundary range of the target field is calculated, and a preset redundant margin is added to the periphery of the display boundary range to obtain the effective display area of ​​the target field; one of the four corner points of the computer screen is selected as a positioning reference point, and the coordinates of the positioning reference point are used in combination with the effective display area to generate a display matrix of the target field, wherein the display matrix contains the position information and pixel values ​​of the target field; the display matrix is ​​output to the computer screen to complete the positioning display of the target field, and the display result is saved as the target image.

[0013] The above technical solution, which adopts a precise positioning mechanism based on the geometric properties of text and a dynamic adjustment strategy for the display matrix, can achieve precise positioning and adaptive display of the original image of text information in the CAD file, effectively solving technical problems such as large deviation in text display position, unreasonable display size, and blurred display boundaries in the existing technology, thereby achieving high-quality visual presentation of text information. This solution ensures the accuracy of the display position through precise positioning of text coordinates, achieves adaptive adjustment of the display size through dynamic calculation of mapping relationships, ensures the rationality of the display range through intelligent calculation of display boundaries, and achieves standardized display of text information through the coordination of positioning reference points and the display matrix, ultimately achieving the technical effects of improving text display quality, optimizing display effects, and enhancing user experience.

[0014] Optionally, the step of calling the first language model to perform text recognition on each image in the image file and arranging the recognition results to form a text list specifically includes: The first language model is called to perform text recognition on each image in the image file to obtain a preliminary recognition result; the preliminary recognition result is subjected to contextual intelligent error correction based on the domain knowledge graph to obtain a final recognition result; the first language model is called to analyze the text coherence of the final recognition result, and the text is arranged to generate a final text; the final text is organized according to a predetermined format to obtain the text list.

[0015] The above-mentioned technical solution, which utilizes a large language model-based intelligent text recognition mechanism and a domain knowledge graph-based contextual error correction strategy, achieves high-precision recognition and intelligent error correction of text in images. This effectively addresses technical issues such as low text recognition accuracy, difficulty in recognizing specialized terminology, and lack of contextual relevance in existing technologies, thereby achieving high-quality output of text recognition results. This solution achieves preliminary text recognition through a large language model, performs specialized error correction through a domain knowledge graph, ensures the semantic integrity of the text through text coherence analysis, and ensures the uniformity of the output results through standardized organization in a predetermined format. Ultimately, it achieves the technical effects of improving text recognition accuracy, enhancing the usability of recognition results, and improving processing efficiency.

[0016] Optionally, the step of calling the second language model to obtain summary information of the target CAD file according to the preset prompt template and the text list specifically includes: Obtain the preset prompt template; input the preset prompt template and the text list as input information into the second language model to obtain summary information of the target CAD file; and output the summary information of the target CAD file to the user.

[0017] The above technical solution, employing a guidance mechanism based on preset prompt templates and an intelligent processing strategy using a second language model, achieves standardized generation and intelligent extraction of CAD file summary information. This effectively addresses existing technical issues such as inconsistent summary formats, incomplete extracted content, and unstable output quality, thereby achieving high-quality summary output. This solution standardizes the format of input information through preset prompt templates, enables intelligent information processing through a large language model, and ensures the uniformity of summary information through standardized output. Ultimately, this achieves the technical benefits of improving summary generation efficiency, enhancing summary practicality, and optimizing the user experience.

[0018] Optionally, after the steps of calling a large language model according to a preset large language model calling strategy, and calling a second large language model according to a preset prompt template and the text list to obtain summary information of the target CAD file, the method further includes: The abstract information is subjected to subject feature extraction, wherein the subject feature is used to characterize key information points of the technical content and design intent of the abstract information; a classification catalog is established according to preset professional field classification rules, and the abstract information is classified and stored according to the subject feature.

[0019] The above technical solution, which employs a topic-based information extraction mechanism and a specialized field classification storage strategy, achieves intelligent organization and systematic management of summary information. This effectively addresses existing technical issues such as chaotic information classification, irrational storage structure, and low retrieval efficiency, thereby enabling efficient management and rapid retrieval of summary information. This solution accurately identifies key information through topic feature extraction, ensures a rational storage structure through specialized field classification, and improves information retrieval efficiency through a classified storage mechanism. Ultimately, it achieves the technical benefits of optimizing information management, improving retrieval performance, and enhancing system practicality.

[0020] In a second aspect, an embodiment of the present application provides a device for automatically extracting summary information from CAD files, the device comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory being used to store computer program code, the computer program code comprising computer instructions, the one or more processors calling the computer instructions to enable the device for automatically extracting summary information from CAD files to execute the method described in the first aspect and any possible implementation of the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer-readable storage medium comprising instructions. When the instructions are executed on a device for automatically extracting summary information of a CAD file, the device for automatically extracting summary information of a CAD file executes the method described in the first aspect and any possible implementation of the first aspect.

[0022] In a fourth aspect, an embodiment of the present application provides a computer program product comprising instructions. When the above-mentioned computer program product is run on a CAD file summary information automatic extraction device, the above-mentioned CAD file summary information automatic extraction device executes the method described in the first aspect and any possible implementation method of the first aspect.

[0023] It is understood that the automatic CAD file summary information extraction device provided in the second aspect, the computer storage medium provided in the third aspect, and the computer program product provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects achieved by these methods can be referenced to the beneficial effects of the corresponding methods and will not be further elaborated here.

[0024] In summary, one or more technical solutions provided in this application have at least the following technical effects or advantages: 1. CAD file parsing and development tools are used to accurately locate and extract attributes of text information, obtain actual images of text in CAD files, and then process the actual images through a large language model. Summary information is generated based on preset prompt words, achieving accurate recognition and semantic understanding of text information in CAD files. This effectively solves technical problems in the existing technology of directly extracting text information from CAD files, such as inaccurate extraction results, lack of contextual relevance, and low information extraction efficiency. This allows for the automated, intelligent, and standardized extraction of summary information from CAD files.

[0025] 2. The system adopts a historical search keyword analysis mechanism and a dynamic keyword adjustment strategy based on the third language model to intelligently optimize and supplement the search keywords entered by users, effectively solving technical problems in existing technologies such as single search keywords, low matching accuracy, and inaccurate search results, thereby achieving a significant improvement in search efficiency and accurate matching of search results.

[0026] 3. The information extraction mechanism based on subject features and the professional field classification storage strategy are adopted to realize the intelligent organization and systematic management of summary information, effectively solving the technical problems of chaotic information classification, unreasonable storage structure, low retrieval efficiency and so on in the existing technology, thereby realizing the efficient management and rapid retrieval of summary information. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flow chart of a method for automatically extracting summary information from a CAD file provided in an embodiment of the present application; Figure 2 This is a flow chart of another method for automatically extracting summary information from CAD files provided in an embodiment of the present application; Figure 3 This is a schematic diagram of the physical device structure of the CAD file summary information automatic extraction device in the embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0029] In the description of the embodiments of this application, words such as "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "for example" or "for instance" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "for example" or "for instance" is intended to present the relevant concepts in a concrete manner.

[0030] In the description of the embodiments of the present application, the term "plurality" means two or more. In addition, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of such features. The terms "include," "comprise," "have" and their variations all mean "including but not limited to," unless otherwise specifically emphasized.

[0031] The following is combined with Figure 1 and Figure 2 The embodiments of the present application are described.

[0032] This application provides a method for automatically extracting summary information of CAD files. Figure 1 , Figure 1 This is a flow chart of a method for automatically extracting summary information from a CAD file provided in an embodiment of the present application. The method includes: Step S101: parsing a target CAD file using a CAD file parsing development tool to obtain first text information in the target CAD file, where the first text information includes a first field set and text geometric attribute information corresponding to each field in the first field set, where the text geometric attribute information includes text coordinates, text size, and text rotation angle. Among them, the CAD file parsing development tool refers to a professional software development toolkit used to parse and process CAD file formats; the target CAD file represents a CAD format file that requires text extraction and processing; the first text information refers to the complete text content extracted from the target CAD file; the first field is used to represent each independent text content unit; the text geometric attributes represent a set of parameters that describe the spatial position and morphological characteristics of the text in the CAD file, including the position of the text in the coordinate system (text coordinates), display size (text size), and rotation direction (text rotation angle).

[0033] This step, performed during system initialization, extracts all text information and its attributes from the CAD file. Specifically, the system first calls the CAD file parsing development tool to open the target CAD file. It then sequentially traverses each layer and entity within the file, identifying the text objects within it. It then extracts the content and corresponding geometric attribute information for each text object, including its precise position in the CAD coordinate system (text coordinates), font size (text size), and text rotation angle (text rotation angle), ultimately forming a data set containing the complete text information.

[0034] In some embodiments, text information extraction and attribute parsing can be achieved through a variety of methods: optionally, by calling the AutoCADObjectARX development interface to obtain a file handle, traversing text entity objects, reading text content and attributes, and saving the results to a data structure; optionally, using an open source CAD parsing library such as LibreCAD to parse DXF format files, identify TEXT type entities, extract text content and attribute parameters, and generate a text information set. It is understood that other CAD file parsing methods can also be used to achieve text information extraction, which is not limited here.

[0035] Step S102: Search the first text information according to the current search keyword to obtain second text information, where the second text information includes a second field set and the text geometric attribute information corresponding to each field in the second field set, wherein the second field set includes some fields in the first field set; Among them, the search keyword refers to the keyword or phrase entered by the user to find specific content in the CAD file; the first text information refers to the collection of all text content and its attribute information parsed from the CAD file; the second text information refers to the subset of text content related to the search keyword after search screening; the second field set is used to represent the group of text fields that meet the search conditions; text geometric attribute information refers to the data set that describes the spatial position and display characteristics of text in the CAD file.

[0036] This step, performed after the CAD file is parsed and the first text information is obtained, is used to filter out content relevant to the search request from all the first text information. Specifically, the system first receives the search keyword entered by the user, then performs a keyword match on each field in the first text information. The successfully matched fields and their corresponding geometric attribute information are extracted to form the second text information, which is then accurately positioned and displayed.

[0037] In some embodiments, the keyword search and information extraction process can be implemented in a variety of ways: Optionally, a fuzzy matching method is used: first, the search keyword is segmented; then the similarity between each segmented word and each field in the first text information is calculated; finally, fields with a higher matching degree and their attribute information are screened out based on a similarity threshold.

[0038] Optionally, an exact matching method is used: first, a standardized expression of the search keyword is established; then, the fields in the first text information are normalized; and finally, the fields containing the keyword and their attribute information are extracted through a complete matching method.

[0039] It is understandable that other retrieval algorithms and matching strategies may also be used to extract and filter key information, which is not limited here.

[0040] Step S103: positioning and adjusting each field in the second field set based on the text geometric attribute information corresponding to each field in the second field set using the CAD file parsing and development tool to obtain an image file, wherein the image file includes an image corresponding to each field in the second field set; Among them, the second field set represents the text field group after retrieval and screening; text geometric attribute information refers to the data set that describes the spatial position characteristics of the text in the original CAD file, including text coordinates, text size and rotation angle; the text area represents the specific location range of each text field in the original CAD file; positioning adjustment display refers to the process of mapping the text area in the original CAD file to the computer display interface; the image file includes a display screenshot of each text area after positioning adjustment.

[0041] This step, performed after acquiring the second text information, precisely maps the text area from the original CAD file to the computer display interface. Specifically, the system first locates the precise area of ​​each text field in the original CAD file based on the text's geometric attributes. It then calculates the positional parameters and display properties of these areas when mapped to the display interface. Using a CAD file parsing tool, the system individually adjusts the display of each text area based on the calculated results, and saves each display result as a separate image, ensuring that each text field accurately reproduces its spatial layout characteristics from the original CAD file. By independently locating and displaying each text field in the original CAD file, each text area is fully captured, avoiding issues such as text truncation, omissions, or overlap that can occur with traditional methods. This precise area positioning method ensures the integrity and accuracy of the extracted information. Furthermore, by retaining the geometric attributes (coordinates, size, and angles) of the text in the original CAD file and performing corresponding transformations during display, the displayed result accurately reproduces the spatial layout characteristics of the text in the original design. This ensures more accurate results for subsequent text processing.

[0042] In some embodiments, the positioning adjustment display of the text area can be achieved in a variety of ways: Optionally, a regional positioning method is used: first, the target text area is determined in the original CAD file according to the text coordinates; then the scaling ratio and rotation angle of the area in the display interface are calculated; finally, the display parameters of the text area are adjusted according to the calculation results and captured and saved as an image.

[0043] Optionally, a matrix transformation method is used: first, a spatial transformation matrix of the text area is constructed; then, display transformation parameters are calculated based on the text geometric attribute information; finally, the precise position and shape of the text area in the display interface are determined through matrix operations, and the display result is saved as an image.

[0044] It is understandable that other area positioning and display transformation methods may also be used to achieve accurate mapping of the text area in the original CAD file to the display interface, which is not limited here.

[0045] Step S104: calling the first language model to perform text recognition on each image in the image file, and arranging the recognition results to form a text list; Among them, the first language model represents the artificial intelligence model used for text recognition, including but not limited to specially trained artificial intelligence models or artificial intelligence models with relevant functions on the market; text recognition refers to the process of converting text in an image into editable text; arrangement processing is used to represent the standardization and organization of recognition results; text list refers to the processed structured text data.

[0046] This step is performed after obtaining the image file containing the image corresponding to each field in the second field set. The main purpose is to convert the text information stored in the image form into a structured text format that can be processed by the computer. Specifically, the system first calls the first language model, which has advanced image recognition and text extraction capabilities. It scans the text area and recognizes the content of each image in the image file in turn. During the recognition process, the deep learning characteristics of the model are used to accurately extract the text content in the image; after that, the system will perform unified format normalization processing on all recognized text contents, including but not limited to character encoding unification, special symbol processing, line break normalization, etc. At the same time, the text is arranged and organized in a reasonable order according to the spatial layout relationship of the text in the original CAD file; finally, the system organizes the processed text content into a standardized text list according to a predefined data structure. The list not only retains the semantic information of the original text, but also maintains the logical association relationship between the text, providing a structured data basis for subsequent summary generation.

[0047] Step S105 , based on a preset prompt template and the text list, calling a second language model to obtain summary information of the target CAD file, wherein the preset prompt template is used to instruct to extract information associated with the search keyword from the text list and output it in a predetermined format.

[0048] Among them, the preset prompt template represents a standardized instruction format used to guide information extraction; the second largest language model refers to an artificial intelligence model used to generate summaries, including but not limited to specially trained artificial intelligence models or artificial intelligence models with relevant functions on the market, and can be the same model as the first largest language model; the summary information is used to represent a general description of the core content of the CAD file; the predetermined format refers to a standardized summary output form.

[0049] This step, performed after obtaining the text list, extracts key information from the text content and generates a standardized summary. Specifically, the system takes a preset prompt template and the text list as input. Using the second language model, it analyzes the relevance of the text content to the search keywords, extracts relevant information, organizes it according to a predefined format, and ultimately generates a structured summary.

[0050] In some embodiments, summary generation can be achieved in a variety of ways: Optionally, a template filling method is used: first, the structure of the preset prompt template is parsed; then, the corresponding key information is extracted from the text list; finally, the extracted information is organized according to the template format to generate a summary.

[0051] Optionally, a semantic analysis method is used: first, semantic understanding is performed on the text list; then important information related to the search keywords is identified; and finally, summary content is generated through natural language generation technology.

[0052] It is understandable that other information extraction and summary generation methods may also be used to achieve a summary of the CAD file content, which is not limited here.

[0053] Through the above technical solution, a CAD file parsing development tool is used to accurately locate and extract attributes from text information, obtain actual images of the text in the CAD file, then process the actual images through a large language model, and generate summary information based on preset prompt words. This achieves accurate recognition and semantic understanding of the text information in the CAD file, effectively solving the technical problems of direct CAD file text information extraction in the existing technology, such as inaccurate extraction results, lack of contextual relevance, and low information extraction efficiency. It also achieves automated, intelligent, and standardized extraction of CAD file summary information. This solution ensures the accuracy of text information extraction through precise positioning of text geometric attribute information, ensures the integrity of the extracted content through step-by-step image recognition and semantic understanding, and ensures the uniformity of the output format through standardized guidance of preset prompt templates. Ultimately, it achieves the technical effects of improving the efficiency of CAD file information extraction, reducing manual processing costs, and enhancing the usability of extraction results.

[0054] The following is a more detailed description of the process of the method provided by this implementation. Figure 2, is another flow chart of the method for automatically extracting summary information of CAD files in an embodiment of the present application.

[0055] S201, parsing a target CAD file using a CAD file parsing development tool to obtain first text information in the target CAD file, where the first text information includes a first field set and text geometric attribute information corresponding to each field in the first field set, where the text geometric attribute information includes text coordinates, text size, and text rotation angle; For details, please refer to step S101, which will not be described again here.

[0056] S202, analyzing historical search keywords using a third language model, and adjusting current search keywords; This includes but is not limited to specially trained artificial intelligence models or artificial intelligence models with relevant functions on the market, and can be the same model as the first language model or the second language model; retrieval keyword analysis represents the process of semantic understanding and relevance evaluation of keywords.

[0057] This step is executed after receiving the user's search keyword input and is used to improve the accuracy and efficiency of the search. Specifically, the system first loads the third language model and the historical search record database to perform semantic analysis on the currently entered search keyword. Simultaneously, the system references the historical search record database to evaluate the keyword's search performance. Based on the evaluation results, the system supplements, replaces, or expands the keyword, ultimately obtaining optimized search keywords to improve the accuracy of subsequent searches.

[0058] In some embodiments, intelligent adjustment of search keywords can be achieved in a variety of ways: Optionally, a semantic expansion method is used: first, semantic features of the current search keyword are extracted; then, the similarity with the historical efficient search keyword is calculated based on the word vector model; finally, appropriate supplementary keywords are selected for expansion based on the similarity.

[0059] Optionally, a statistical analysis method is used: first, the usage frequency and success rate of keywords in historical searches are counted; then a keyword optimization rule model is established; finally, the current search keywords are adjusted and optimized according to the rules.

[0060] It is understandable that other keyword analysis and optimization methods may also be used to improve search results, which are not limited here.

[0061] In addition, this application provides a more specific method for analyzing historical search keywords through the third language model and adjusting the current search keywords, as follows: A10, calling the third language model to perform cluster analysis on the historical search keywords to obtain multiple keyword categories, each keyword category including a set of keywords with similar semantics; Among them, cluster analysis is used to represent a data processing method for classifying semantically similar keywords; keyword categories represent keyword groups with similar semantic features; and semantically similar keyword sets refer to groups of keywords that are closely related in meaning.

[0062] This step, executed when the system activates the keyword optimization function, is used to build a semantic classification system for keywords. Specifically, the system first loads the third-largest language model and the historical search keyword database, using the model's semantic understanding capabilities to perform feature extraction and semantic analysis on historical keywords. Then, the third-largest language model automatically clusters keywords based on semantic similarity, automatically grouping keywords with similar semantic features into the same category. This ultimately forms a multi-level keyword classification system, providing a foundation for subsequent keyword matching.

[0063] A20, performing similarity matching between the current search keyword and the plurality of keyword categories to determine a target keyword category; Among them, similarity matching means calculating the semantic similarity between the current search keyword and the existing category; the target keyword category refers to the keyword classification with the closest semantics to the current search keyword; similarity calculation means the mathematical method of evaluating the semantic similarity of texts.

[0064] This step, executed after obtaining the user-entered search keyword, determines the semantic category to which the current search request belongs. Specifically, the system first extracts semantic features from the user-entered search keyword. It then calculates the semantic similarity between the keyword and each established keyword category. Using a pre-defined matching algorithm, it selects the category with the highest similarity as the target keyword category, providing a precise semantic scope for subsequent keyword optimization.

[0065] A30, based on the target keyword category and the industry terminology knowledge base, supplementing and optimizing the search keyword input by the user to generate the adjusted search keyword; Among them, the industry terminology knowledge base refers to a data set containing professional vocabulary and term definitions; keyword supplementation refers to adding relevant professional terms to the original keywords; keyword optimization is used to represent the process of improving and perfecting keywords; and the adjusted search keywords represent the final search conditions after supplementation and optimization.

[0066] This step, performed after the target keyword category is determined, is used to improve the professionalism and completeness of the search keywords. Specifically, the system first accesses the industry terminology knowledge base to extract professional terms and synonyms related to the target keyword category. It then intelligently combines and adjusts these professional terms with the user-entered search keywords based on pre-set optimization rules, ultimately generating an optimized search keyword that retains the original search intent while maintaining professional integrity.

[0067] S203, searching the first text information according to the adjusted search keyword to obtain second text information matching the adjusted search keyword; Among them, the adjusted search keywords represent the search conditions optimized by the third language model; the first text information refers to the set of all text contents parsed from the CAD file; the search matching represents the process of judging the relevance between the keywords and the text contents; the second text information is used to represent the successfully matched text contents and their attribute information set.

[0068] This step is performed after the search keyword adjustment is complete, and is used to achieve accurate text information retrieval. Specifically, the system uses the adjusted and optimized search keyword as the search condition, comprehensively scans and matches the first text information parsed from the CAD file, determines the text content related to the search condition through similarity calculation, and extracts the successfully matched content along with its attribute information to form the second text information, providing an accurate data foundation for subsequent processing.

[0069] In some embodiments, accurate retrieval of text information can be achieved through various methods: Optionally, a multi-dimensional matching method is used: first, a feature vector of the search keyword is constructed; then, the multi-dimensional similarity with the text content is calculated; and finally, the matching results are filtered according to a similarity threshold.

[0070] Optionally, a hierarchical filtering method is used: first, coarse-grained keyword matching is performed; then, a refined comparison is performed on the initial screening results; and finally, the final matching content is determined through comprehensive scoring.

[0071] It is understandable that other search algorithms and matching strategies may also be used to achieve accurate extraction of text information, which is not limited here.

[0072] S204, using the CAD file parsing development tool, locating the text area corresponding to the target field according to the text coordinates corresponding to the target field; Among them, CAD file parsing development tools refer to professional software tools used to process CAD files; text coordinates refer to numerical parameters that describe the position of text in a CAD file; target fields represent specific text content that needs to be positioned; and text areas are used to represent the complete display range that includes the target field.

[0073] This step, performed after the target field to be processed is determined, accurately locates the display position of the text content. Specifically, the system calls a CAD file parsing development tool to read the coordinate information of the target field. Based on this coordinate information, the system determines the precise position of the text in the original CAD file. It then calculates the complete display area containing the text, ensuring that the text content is fully and accurately presented during subsequent display.

[0074] In some embodiments, precise positioning of the text area can be achieved through various methods: Optionally, a coordinate mapping method is used: first, the text coordinate data is parsed; then the space occupied by the text is calculated; and finally, the complete display area boundary is determined.

[0075] Optionally, a regional scanning method is used: first, the scanning range is set with the coordinate as the center; then the scanning area is gradually expanded until the entire text is covered; finally, the regional boundary is optimized to determine the final display range.

[0076] It is understandable that other coordinate processing and area division methods may also be used to achieve precise positioning of the text area, which is not limited here.

[0077] S205, calculating the screen display size of the target field according to the mapping relationship between the text size corresponding to the target field in the text area and screen pixels; Among them, text size refers to the font size parameter of the target field in the original CAD file; screen pixel refers to the smallest display unit of the display device; the mapping relationship is used to represent the conversion rule between the physical size of text and display pixels; screen display size refers to the actual display size of text on the display device.

[0078] This step is performed after the text area is positioned and is used to determine the appropriate display size of the text on the display device. Specifically, when calculating text display size, the system first obtains the specific font size information of the target field from the original CAD file (for example, a font height of 3.5 mm), and simultaneously obtains the preset standard mapping rule (that is, 5 mm text corresponds to 80 screen pixels). Then, based on this standard mapping rule, the system calculates the unit conversion ratio (that is, 1 mm text should be displayed as 16 screen pixels), and then applies this ratio to the actual font size of the current target field for proportional calculation (for example, 3.5 mm text should be displayed as 3.5×16=56 screen pixels). Furthermore, the system will consider the physical characteristics and display requirements of the display device and make necessary corrections to the calculation results, such as rounding up the calculation results to ensure pixel alignment (adjusting 56.0 pixels to 56 pixels), or making corresponding adjustments when the calculation result is smaller than the minimum recognizable size (for example, if the calculation result is less than 40 pixels, it will be uniformly adjusted to 40 pixels to ensure the recognition effect of the computer vision system). Finally, the text display size is obtained that is suitable for clear presentation on the display device and convenient for subsequent recognition processing, ensuring that the text display effect is clear and beautiful while maintaining the proportional relationship with the original design.

[0079] In some embodiments, the text display size can be calculated in a variety of ways: Optionally, a proportional mapping method is used: first, the resolution parameters of the display device are obtained; then, the conversion ratio between the text size and pixels is calculated; and finally, the final display size is determined according to the ratio.

[0080] Optionally, an adaptive calculation method is used: first, the space limitations of the display area are analyzed; then, adaptation parameters are calculated based on the text size; and finally, the optimal display size is dynamically adjusted.

[0081] It is understandable that other size calculation and display optimization methods may also be used to achieve reasonable display of text, which is not limited here.

[0082] S206: Calculate the display boundary range of the target field based on the screen display size and the text rotation angle corresponding to the target field, and add a preset redundant margin to the periphery of the display boundary range to obtain an effective display area of ​​the target field; Among them, the screen display size refers to the pixel size occupied by the text on the display device; the text rotation angle refers to the degree of rotation of the text relative to the horizontal direction; the display boundary range refers to the minimum rectangular area that contains the complete text content; the redundant margin refers to the additional blank area added outside the boundary to ensure the integrity of the text display; the effective display area is used to represent the complete space range ultimately used for display.

[0083] This step, performed after calculating the text's screen display size, determines the text's complete display range. Specifically, the system first calculates the actual space occupied by the rotated text using trigonometric functions based on the text's display size and rotation angle. This determines the minimum rectangular boundary that can fully encompass the text, then uniformly adds a preset redundant margin (usually tens of pixels) around this boundary to form the final effective display area, ensuring the integrity and legibility of the text display.

[0084] S207, selecting one of the four corner points of the computer screen as a positioning reference point, and using the coordinates of the positioning reference point in combination with the effective display area to generate a display matrix of the target field, wherein the display matrix includes the position information and pixel values ​​of the target field; Among them, the corner points represent the positions of the four vertices of the computer screen; the positioning reference point refers to the reference position selected for text positioning; the display matrix represents a two-dimensional array that describes the text display position and content; the position information refers to the coordinate data of the text on the screen; the pixel value is used to represent the color and transparency information of the display content.

[0085] This step, performed after the effective display area is determined, generates the specific display data for the text. Specifically, the system first selects one of the four corners of the screen—the upper left, upper right, lower left, and lower right—as a positioning reference point. Then, based on the previously calculated effective display area size, it calculates the specific position of the text display through coordinate transformation. Finally, the text content is converted into a pixel matrix, forming a display matrix containing the complete display information.

[0086] S208, outputting the display matrix to the computer screen, completing the positioning display of the target field, and saving the display result as the target image; Among them, display matrix output means transmitting display data to the display device; positioning display refers to accurately presenting text content at a specified location; target image represents the saved display result file; and the computer screen represents the output device for displaying text content.

[0087] This step, performed after the display matrix is ​​generated, is used to display the text and save the results. Specifically, the system first transfers the display matrix data to the display buffer. The display driver then accurately renders the text on the computer screen, ensuring that the text position, size, and rotation angle match the calculated results. The resulting display is then converted into an image format and saved as input for subsequent text recognition.

[0088] S209, calling the first language model to perform text recognition on each image in the image file to obtain a preliminary recognition result; Among them, the first language model represents an intelligent model specifically used for text recognition; text recognition refers to converting text in an image into text that can be processed by a computer; the preliminary recognition result represents the original text data output by the model; and the image file is used to represent the image data set to be recognized.

[0089] This step, executed after the text is displayed and the image is saved, converts the image text into text data. Specifically, the system first loads the trained language model, then processes each image in the image file in turn, using the model's text recognition capabilities to convert the text in the image into a computer-processable text format. This yields preliminary text recognition results, providing foundational data for subsequent text processing.

[0090] S210, performing contextual intelligent error correction on the preliminary recognition result based on the domain knowledge graph to obtain a final recognition result; Among them, the domain knowledge graph represents the knowledge system and concept association network of a specific professional field; the preliminary recognition result refers to the original text content output by the text recognition model; contextual intelligent error correction refers to the process of text correction based on context and professional knowledge; the final recognition result is used to represent the accurate text information after error correction; intelligent error correction refers to the process of using artificial intelligence technology to detect and correct text errors.

[0091] This step is performed after obtaining preliminary text recognition results and is used to improve text recognition accuracy. Specifically, the system first loads a domain-specific knowledge graph database. It then performs contextual analysis on each character in the preliminary recognition results. By matching and comparing these characters with professional terms and common expressions in the knowledge graph, it identifies potential recognition errors and intelligently corrects them based on context and professional knowledge, ultimately achieving accurate text recognition results.

[0092] S211, calling the first language model to analyze the text coherence of the final recognition result, and performing text arrangement to generate a final text; Among them, text coherence refers to the semantic and logical fluency of the text content; text arrangement means the reasonable organization of the text order; and the final text is used to represent the complete and organized text content.

[0093] This step is performed after text correction to ensure the semantic coherence of the text. Specifically, the system uses the first language model to perform in-depth semantic analysis on the corrected text, assessing the logical connections between words. It then reorders and organizes the text based on the language characteristics of professional documents, ensuring that the generated text maintains coherence and professionalism in both structure and content.

[0094] S212, arranging the final text according to a predetermined format to obtain the text list; Among them, the predetermined format refers to the prescribed text organization form and display style; the final text refers to the text content that has completed coherence processing; and the text list is used to represent structured text data organized according to a specific format.

[0095] This step, performed after the text coherence analysis is complete, standardizes the presentation of the text. Specifically, the system structures the coherently processed text according to predefined text formatting requirements, including paragraph division, indentation, and punctuation, ultimately generating a standardized text list that meets the predefined formatting requirements.

[0096] S213, obtaining the preset prompt template; Among them, the preset prompt template represents a standardized instruction format for guiding information extraction and organization; This step, executed after the text list is generated, prepares the guiding rules for subsequent information extraction. Specifically, the system reads a predefined prompt template from the configuration database. This template contains guiding content such as information extraction rules, keyword recognition methods, and output formats, providing a standardized processing basis for subsequent text processing.

[0097] S214, inputting the preset prompt template and the text list as input information into the second language model to obtain summary information of the target CAD file; The summary information refers to a concise summary of the core content of the target CAD file.

[0098] This step, performed after the prompt template and text list are prepared, generates a summary of the CAD file's contents. Specifically, the system combines the preset prompt template and standardized text list as input data. Using the second language model, the system performs deep semantic understanding and information extraction. It then intelligently analyzes and summarizes the text according to the requirements defined in the template, ultimately generating a summary that accurately reflects the core content of the CAD file.

[0099] S215, outputting summary information of the target CAD file to the user; Among them, output refers to the process of displaying the processing results to the user; the user refers to the operator of the system or the recipient of information.

[0100] This step is performed after the summary information is generated and is used to display the processing results to the user. Specifically, the system formats the generated summary information through the user interface to ensure readability and practicality. It may also provide additional interactive functions so that users can easily view, copy, or further process the summary content.

[0101] In some embodiments, the output of the summary can be achieved in multiple ways: Optionally, use a visual display method: first design a friendly display interface; then format the summary content; finally, clearly display it to the user through the interface.

[0102] Optionally, an interactive output method is used: first, an interactive operation interface is constructed; then, a variety of viewing options are provided; and finally, personalized operations are supported for users.

[0103] It is understandable that other display modes and interaction methods may be used to achieve effective output of summary information, which is not limited here.

[0104] S216, extracting subject features from the summary information, wherein the subject features are used to characterize key information points of the technical content and design intent of the summary information; Among them, the subject features represent the characteristic identifiers of the core technical content and design intent in the abstract information, such as: technical classification features (such as "mechanical transmission", "hydraulic system", "electrical control"), functional features (such as "automatic feeding", "constant pressure control", "temperature compensation"), structural features (such as "box structure", "sliding mechanism", "sealing device"), performance features (such as "high precision", "low noise", "energy saving and environmental protection"), and application scenario features (such as "CNC machining", "automatic assembly", "online detection"); technical content refers to the specific technical implementation details in the CAD file; design intent is used to express the designer's purpose and considerations; key information points represent content elements that are particularly important for understanding the design.

[0105] This step, executed after the summary is output, deeply analyzes the technical features of the summary content. Specifically, the system first identifies and categorizes the technical terms in the summary information. It then extracts feature information from multiple dimensions, including technical fields, functional implementation, structural design, performance indicators, and application scenarios. It also analyzes the relationships between these features, ultimately forming a multi-layered thematic feature network that comprehensively reflects the technical key points and innovative ideas of the design.

[0106] S217, establishing a classification directory according to preset professional field classification rules, and classifying and storing the summary information according to the subject features; Among them, professional field classification rules represent the classification standards used for systematic management of technical documents, such as classification by technical field (mechanical design, electrical control, software system, etc.), functional type (transmission system, control system, detection system, etc.), and application scenario (industrial production, medical equipment, agricultural machinery, etc.); classification directory refers to the document storage system constructed according to professional rules; classification storage is used to represent the process of systematically archiving documents according to specific categories.

[0107] This step, performed after obtaining complete subject features, enables standardized storage of summary information. Specifically, the system first establishes a multi-level classification directory structure based on industry standards and professional specifications. It then analyzes the correspondence between subject features and classification rules, selecting the most appropriate storage category for the summary information. Finally, the summary information, along with its subject features, is stored in the corresponding classification directory, enabling intelligent document classification management.

[0108] The following describes the automatic extraction device for CAD file summary information in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , which is a schematic diagram of the physical device structure of the CAD file summary information automatic extraction device in an embodiment of the present application.

[0109] It should be noted that Figure 3 The structure of the device for automatically extracting summary information of a CAD file is merely an example and should not limit the functions and scope of use of the embodiments of the present invention.

[0110] like Figure 3 As shown, the automatic CAD file summary information extraction system includes a CPU 301, which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 302 or programs loaded from a storage unit 308 into a random access memory (RAM) 303, such as the methods described in the above embodiments. RAM 303 also stores various programs and data required for system operation. CPU 301, ROM 302, and RAM 303 are interconnected via a bus 304. An I / O interface 305 is also connected to bus 304.

[0111] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, push button switches, and the like; an output section 307 including a liquid crystal display (LCD), an audio output device, indicator lights, and the like; a storage section 308 including a hard disk and the like; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. Removable media 311, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 310 as needed, so that computer programs read from the removable media can be installed in the storage section 308 as needed.

[0112] In particular, according to an embodiment of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, an embodiment of the present invention includes a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 309 and / or installed from the removable medium 311. When the computer program is executed by the CPU 301, the various functions defined in the present invention are performed.

[0113] It should be noted that specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0114] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings.

[0115] Specifically, the automatic extraction device for CAD file summary information of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, the automatic extraction method for CAD file summary information provided by the above embodiment is implemented.

[0116] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the device for automatically extracting CAD file summary information described in the above embodiments, or may exist independently and not be incorporated into the device. The storage medium carries one or more computer programs, which, when executed by a processor of the device for automatically extracting CAD file summary information, enable the device to implement the method for automatically extracting CAD file summary information provided in the above embodiments.

[0117] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0118] As used in the above embodiments, the term “when” may be interpreted to mean “if” or “after” or “in response to determining that” or “in response to detecting that”, depending on the context. Similarly, the phrases “upon determining that” or “if (stated condition or event) is detected” may be interpreted to mean “if determining that” or “in response to determining that” or “upon detecting (stated condition or event)” or “in response to detecting (stated condition or event)”, depending on the context.

[0119] Those skilled in the art will appreciate that all or part of the process steps in the above-described method embodiments can be implemented by a computer program instructing the relevant hardware. The program can be stored in a computer-readable storage medium, and when executed, the program can include the process steps in the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for automatically extracting summary information from a CAD file, characterized in that: include: Parsing the target CAD file using a CAD file parsing development tool to obtain first text information in the target CAD file, the first text information including a first field set and text geometric attribute information corresponding to each field in the first field set, the text geometric attribute information including text coordinates, text size, and text rotation angle; Searching the first text information according to a current search keyword to obtain second text information, where the second text information includes a second field set and the text geometric attribute information corresponding to each field in the second field set, wherein the second field set includes some fields in the first field set; According to the text geometric attribute information corresponding to each field in the second field set, positioning and adjusting the display of each field in the second field set using the CAD file parsing and development tool to obtain an image file, wherein the image file includes an image corresponding to each field in the second field set; Calling the first language model to perform text recognition on each image in the image file, and arranging the recognition results to form a text list; According to a preset prompt template and the text list, the second largest language model is called to obtain summary information of the target CAD file, wherein the preset prompt template is used to instruct to extract information associated with the search keyword from the text list and output it in a predetermined format.

2. The method according to claim 1, characterized in that The step of searching the first text information according to the current search keyword to obtain the second text information specifically includes: Analyze historical search keywords through the third language model and adjust the current search keywords; The first text information is searched according to the adjusted search keyword to obtain second text information matching the adjusted search keyword.

3. The method according to claim 2, characterized in that The step of analyzing historical search keywords by using the third language model and adjusting the current search keywords specifically includes: Calling the third language model to perform cluster analysis on the historical search keywords to obtain a plurality of keyword categories, each of the keyword categories including a set of keywords with similar semantics; Perform similarity matching between the current search keyword and the plurality of keyword categories to determine a target keyword category; Based on the target keyword category and the industry terminology knowledge base, the search keyword input by the user is supplemented and optimized to generate the adjusted search keyword.

4. The method according to claim 1, wherein The step of positioning, adjusting, and displaying each field in the second field set using the CAD file parsing and development tool according to the text geometric attribute information corresponding to each field in the second field set to obtain an image file specifically includes: The target field is any field in the second field set. For the target field, a corresponding target image is obtained in the following manner. The image file includes the target image: Using the CAD file parsing development tool, locating the text area corresponding to the target field according to the text coordinates corresponding to the target field; Calculating the screen display size of the target field according to a mapping relationship between a text size corresponding to the target field in the text area and screen pixels; Calculating a display boundary range of the target field based on the screen display size and the text rotation angle corresponding to the target field, and adding a preset redundant margin to the periphery of the display boundary range to obtain an effective display area of ​​the target field; Selecting one of the four corner points of the computer screen as a positioning reference point, and using the coordinates of the positioning reference point in combination with the effective display area to generate a display matrix of the target field, wherein the display matrix includes the position information and pixel values ​​of the target field; The display matrix is ​​output to the computer screen to complete the positioning display of the target field, and the display result is saved as the target image.

5. The method according to claim 1, characterized in that The step of calling the first language model to perform text recognition on each image in the image file and arranging the recognition results to form a text list specifically includes: Calling the first language model to perform text recognition on each image in the image file to obtain a preliminary recognition result; Perform contextual intelligent error correction on the preliminary recognition result based on the domain knowledge graph to obtain the final recognition result; Calling the first language model to analyze the text coherence of the final recognition result, and performing text arrangement to generate a final text; The final text is sorted according to a predetermined format to obtain the text list.

6. The method according to claim 1, characterized in that The step of calling the second language model to obtain summary information of the target CAD file according to the preset prompt template and the text list specifically includes: Obtaining the preset prompt template; Input the preset prompt template and the text list as input information into the second language model to obtain summary information of the target CAD file; The summary information of the target CAD file is output to the user.

7. The method according to claim 1, characterized in that After the step of calling the second language model to obtain summary information of the target CAD file according to the preset prompt template and the text list, the method further includes: Extracting subject features from the abstract information, wherein the subject features are used to characterize key information points of the technical content and design intent of the abstract information; A classification catalog is established according to preset professional field classification rules, and the summary information is classified and stored according to the subject features.

8. A device for automatically extracting summary information from CAD files, characterized in that: The device for automatically extracting summary information from CAD files includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the device for automatically extracting summary information from CAD files to execute the method described in any one of claims 1 to 7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instruction is executed on a device for automatically extracting summary information of a CAD file, the device is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on a device for automatically extracting summary information from a CAD file, the device is enabled to perform the method according to any one of claims 1 to 7.

Images